Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
ELECTROMAGNETIC SEISMOLOGY VIBRATOR SYSTEMS AND
METHODS
CROSS-REFERENCE TO RELATED APPLICATION
[0001] This application claims benefit of and priority to U.S. provisional
patent
application no. 61/160,405 filed March 16, 2009, which is fully incorporated
herein
by reference and made a part hereof.
BACKGROUND
[0002] Vibroseis, as known in the art, is a method used in exploration
seismology to
propagate energy signals into the earth over an extended period of time as
opposed to
the near instantaneous energy provided by an impulsive source such as
explosives or
weight-drop trucks. Data recorded in this way can be correlated to convert the
extended source signal into an impulse. Generally, the signal has been
generated by a
servo-controlled hydraulic vibrator or shaker unit mounted on a mobile base
unit.
[0003] Techniques for using seismic waves to probe for oil reserves or other
geological
structures and/or anomalies in the earth are becoming quite sophisticated. The
key
requisite for the work is a high fidelity ground shaker whose force,
frequency, and
phase can be controlled precisely. Currently, hydraulic shakers dominate the
industry.
These devices can be truck-mounted for mobility. Generally, with hydraulic
shakers,
small mechanical actuators control high pressure hydraulic fluid flow to apply
sinusoidal compressive forces to a base plate. Large forces are achievable
with this
design. However, there are drawbacks to hydraulic shakers. The use of
hydraulic fluids
can impose environmental concerns regarding leaks and spillage. Furthermore,
hydraulic shakers are relatively slow-responding, which can result in
decreased fidelity
to the input signal.
[0004] Therefore, systems and methods that overcome challenges found in the
art,
some of which are described above, are desired.
1
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
SUMMARY
[0005] Described herein are embodiments of an electromagnetic system that can
be
used to replace the traditional hydraulic oil systems that actuate mass
movement. The
embodiments described herein provide wide frequency range operation, ground
force
application with high fidelity, and low environmental impact. Embodiments
described herein can be used. for seismic exploration and vibroseis
applications,
among other uses.
[0006] Additional advantages will be set forth in part in the description
which follows
or may be learned by practice. The advantages will be realized and attained by
means
of the elements and combinations particularly pointed out in the appended
claims. It
is to be understood that both the foregoing general description and the
following
detailed description are exemplary and explanatory only and are not
restrictive, as
claimed.
BRIEF DESCRIPTION OF THE DRAWINGS
[0007] The accompanying drawings, which are incorporated in and constitute a
part
of this specification, illustrate embodiments and together with the
description, serve to
explain the principles of the methods and systems:
FIG. 1 illustrates an embodiment of a truck-mounted electromagnetic vibrator
system
comprised of a power supply, a control system, and a transducer;
FIG. 2 illustrates one exemplary embodiment of a transducer;
FIG. 3 is a cross section of one embodiment of a transducer;
FIG. 4 further illustrates magnetic forces that cause movement of one
embodiment of
a transducer;
FIG. 5 illustrates one embodiment of an electrical schematic of an inverter
scheme
that can be used to create AC power and control an embodiment of the
electromagnetic
vibrator system;
FIG. 6 shows a pulse-width modulated signal from the inverter can be combined
with
a sawtooth wave to produce a desired sinusoidal wave;
FIGS. 7A and 7B illustrate simulated and actual outputs from the embodiment of
an
inverter as shown in Figure 5;
FIGS. 8A-8C illustrate a schematic of simulation model;
2
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
FIG. 8D illustrates a control diagram;
FIG. 8E illustrates a sweep force control diagram;
FIG. 8F illustrates vibrator electrical waveforms;
FIG. 8G is a graph of spectral power density;
FIG. 8H is a graph of reaction mass motion;
FIGS. 81 and 8J are graphs of random phase ground force comparison;
FIGS. 8K and 8L are illustrations of random phase spectral power comparison;
FIGS. 8M and 8N are illustrations of random amplitude ground force comparison;
FIGS. 80 and 8P are illustrations of random amplitude spectral power
comparison;
FIGS. 8Q, 8R and 8S are illustrations of spectral power plots for fixed
frequency
sweeps;
FIG. 9A is an illustration of an electromagnetic shaker suitable for a large
stroke;
FIG. 9B is an illustration of connecting two coils in anti-parallel will yield
a larger
inductance at higher frequency than shorting one of the two coils;
FIG. 9C is an illustration of geometry for generating lift force from a fixed
field
winding on the steel structure;
FIG. 9D is an illustration of forces generated when the currents are creating
lift for
various positions along the highlighted segment in Figure 10-4;
FIG. 9E is an illustration of forces generated when the currents are creating
a down
force for various positions along the highlighted segment in Figure 10-5;
FIG. 9F is an illustration of force variation with displacement for constant
current;
FIG. 9G is an illustration of baseline design geometry;
FIG. 9H is an illustration of variables to perturb during the optimization;
FIG. 91 is an illustration of a three-dimensional surface is constructed to
link the
optimization variables to the problem geometry;
FIG. 9J illustrates magnetic optimization requires smearing the current over a
region;
FIG. 9K illustrates current density multiplier as a function of (3 where 0
represents the
proportion of the total force load made up by the dc (steady state weight)
component;
FIG. 9L illustrates optimized design dimensions;
FIG. 9M illustrates field plot for the optimized design at maximum excitation;
FIG. 9N illustrates armature coil layout using two parallel circuits;
FIG. 90 illustrates dc lift coils are best positioned in the center of the
armature stack;
FIG. 9P illustrates magnetic field through the middle of the armature gap;
3
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
FIGS. 9Q(a), (b) and (c) illustrate plots of how current density changes with
frequency and in the presence of a copper compensator;
FIG. 9R illustrates power and voltage requirements for thin and thick
compensator
with identical armature coil excitation (225 Hz, maximum excitation);
FIG. 9S illustrates equivalent ladder network used to represent the armature
circuit;
FIG. 9T illustrates performance of the ladder equivalent circuit to model the
armature's power dissipation;
FIG. 9U illustrates transient analysis to compute force with the dc field and
ac
armature excitation;
FIG. 9V illustrates placing two columns magnets on either side of the armature
gap;
FIG. 9W illustrates leakage field away from the armature gap with magnet
excitation;
FIG. 1 OA illustrates heat deflection and glass transition temperatures of
insulation
systems;
FIG. 10B illustrates an insulation system design;
FIG. 11A illustrates an EMV baseline design;
FIG. 11B illustrates a steel reaction mass;
FIG. 11 C illustrates an EMV base plate;
FIG. 11D illustrates field coils;
FIG. 11E illustrates EMV force coils;
FIG. 11F illustrates other EMV Components;
FIG. 11 G illustrates EMV base plate with force coils attached;
FIG. 1 1H illustrates field coils prior to placement in top reaction mass
steel;
FIG. 111 illustrates the outer steel casting prior to copper ring placement;
FIG. 11J illustrates the center steel casting prior to copper ring placement;
FIG. 11K illustrates EMV reaction mass with base plate inserted;
FIG. 11L illustrates EMV base plate guide bearings
FIG. 12A illustrates a block diagram of the control system;
FIG. 13A illustrates a Mertz truck with the new components placed;
FIG. 13B illustrates a Mertz truck showing the top view;
FIG. 13C illustrates a Mertz truck showing the side view;
FIG. 14A illustrates a finite element model of ground and baseplate used in
dynamic
analyses;
FIG. 14B is a detailed view of finite element model of ground and baseplate
used in
4
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
dynamic analyses, with load application areas shown;
FIG. 14C illustrates a finite element mesh of baseplate used in static
analysis;
FIG. 14D illustrates a close up of finite element mesh of baseplate used in
static
analysis;
FIG. 14E illustrates a baseplate lifting up off the ground during upward
application of
the sinusoidal load;
FIG. 14F illustrates the maximum Von Mises stress encountered in the dynamic
analyses - the stress is 14.3 ksi;
FIG. 14G illustrates the maximum Von Mises stress encountered in the upper
cylindrical portion of the baseplate - the stress is 7.5 ksi;
FIG. 14H illustrates maximum Von Mises stress from lateral load in static
analysis;
FIG. 141 illustrates lateral deflection from lateral load;
FIG. 14J illustrates a modified baseplate with stiffening ribs;
FIG. 15A illustrates boiling and freezing temperatures of ethylene glycol
aqueous
solutions (ASHRAE handbook);
FIG. 15B illustrates a field coil cooling design;
FIG. 15C illustrates an armature lift coil and compensating plate cooling
design;
FIG. 15D illustrates a finite-element thermal model mesh for field coils;
FIG. 15E illustrates finite-element thermal model meshing for lift coils and
compensating plates;
FIG. 15F illustrates a field coil steady-state temperature distribution (using
average
coolant temperature);
FIG. 15G illustrates a field coil steady-state temperature distribution (using
exit
coolant temperature);
FIG. 15H illustrates a lift coil steady-state temperature distribution (using
average
coolant temperature);
FIG. 151 illustrates a lift coil steady-state temperature distribution (using
exit coolant
temperature);
FIG. 15J illustrates a compensating plate steady-state temperature
distribution (using
average coolant temperature);
FIG. 15K illustrates a compensating plate steady-state temperature
distribution (using
exit coolant temperature);
FIG. 16A illustrates a magnetic field map in air gap;
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
FIG. 16B illustrates a current check out of a PWM at 10 Hz;
FIG. 16C illustrates a check out of a PWM converter at 50 Hz;
FIG. 17A illustrates springs added between baseplate and reaction mass;
FIG. 17B illustrates a vibrator in full down position;
FIG. 17C illustrates a vibrator in full up position;
FIG. 17D illustrates conductors redesigned with 0.125 in. tubing; and
FIG. 17E illustrates a conductor in Roebel TM configuration.
DETAILED DESCRIPTION
[0008] Before the present methods and systems are disclosed and described, it
is to be
understood that the methods and systems are not limited to specific synthetic
methods, specific components, or to particular compositions. It is also to be
understood that the terminology used herein is for the purpose of describing
particular
embodiments only and is not intended to be limiting.
[0009] As used in the specification and the appended claims, the singular
forms "a,"
"an" and "the" include plural referents unless the context clearly dictates
otherwise.
Ranges may be expressed herein as from "about" one particular value, and/or to
"about" another particular value. When such a range is expressed, another
embodiment includes from the one particular value and/or to the other
particular
value. Similarly, when values are expressed as approximations, by use of the
antecedent "about," it will be understood that the particular value forms
another
embodiment. It will be further understood that the endpoints of each of the
ranges are
significant both in relation to the other endpoint, and independently of the
other
endpoint.
[0010] "Optional" or "optionally" means that the subsequently described event
or
circumstance may or may not occur, and that the description includes instances
where
said event or circumstance occurs and instances where it does not.
[0011] Throughout the description and claims of this specification, the word
"comprise" and variations of the word, such as "comprising" and "comprises,"
means
"including but not limited to," and is not intended to exclude, for example,
other
additives, components, integers or steps. "Exemplary" means "an example of'
and is
not intended to convey an indication of a preferred or ideal embodiment. "Such
as" is
not used in a restrictive sense, but for explanatory purposes.
6
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
[0012] Disclosed are components that can be used to perform the disclosed
methods
and systems. These and other components are disclosed herein, and it is
understood
that when combinations, subsets, interactions, groups, etc. of these
components are
disclosed that while specific reference of each various individual and
collective
combinations and permutation of these may not be explicitly disclosed, each is
specifically contemplated and described herein, for all methods and systems.
This
applies to all aspects of this application including, but not limited to,
steps in
disclosed methods. Thus, if there are a variety of additional steps that can
be
performed it is understood that each of these additional steps can be
performed with
any specific embodiment or combination of embodiments of the disclosed
methods.
[0013] The present methods and systems may be understood more readily by
reference to the following detailed description of preferred embodiments and
the
Examples included therein and to the Figures and their previous and following
description.
[0014] Described herein are systems and methods of an electromagnetic
vibrator. As
shown in Figure 1, in one embodiment an electromagnetic vibrator system can be
comprised of a power supply 202, a control system 204 and a transducer 206.
The
embodiment of Figure 1 is optionally shown in a truck-mounted configuration
and is
further comprised of a prime-mover 208 used for power generation.
[0015] Figure 2 illustrates one exemplary embodiment of a transducer 306,
which is
generally comprised of a reaction mass 302 and a base plate 304.
[0016] Figure 3 is a cross section of one embodiment of a transducer 306. As
shown
in this figure, field windings 402 are embedded within the reaction mass 302.
A field
winding current in the field windings 402 creates a magnetic field that reacts
with a
magnetic field created by an armature current in armature windings 404 that
are
associated with the base plate 304. By varying one or more of the direction,
frequency or magnitude (amplitude) of the field winding current and/or the
armature
- winding current, the base plate and/or the reaction mass can be moved in
upward or
downward directions.
[0017] Figure 4 further illustrates magnetic forces that cause movement of one
embodiment of a transducer 306. As shown in Figure 4, magnetic forces cause
movement of the reaction mass 302 and/or the base plate 304 (only a portion of
the
base plate is shown in Figure 4). The armature winding current (I) 502 creates
a
7
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
magnetic field that interacts with the magnetic field (BR) created by the
field winding
current in the field winding 402. In one embodiment, the armature winding
current
can be a sinusoidally-varying current. The magnetic field created by the
armature
winding current can be controlled by changing one or more of the direction,
amplitude, or frequency of the armature winding current. In one aspect, the
armature
winding comprises Roebel TM windings or other forms of transposed windings, as
are known in the art. In one aspect, the field winding current can be direct
current
(DC).
[0018] In one exemplary embodiment, the shaker can create 60,000 lbf. In one
aspect, it can have a dynamic range of 6 - 150 Hz. In one aspect, the shaken
mass or
reaction mass can comprise 16,000 lbm and the base plate mass can comprise
3,000
lbm.
[0019] - Referring back to Figure 1, in one embodiment a power supply can be
provided. While electrical power supplies as are available from a utility grid
are
contemplated within the scope of embodiments described herein, in one aspect a
prime-mover such as an internal-combustion engine or combustion turbine, for
example, can be provided that drives an electrical generator. Generally, the
generator
produces alternating current (AC) power. In one aspect, the generator can be
single-
phase. In another aspect, the generator can be multiple-phase, such as three-
phase, as
is known in the art.
[0020] Figure 5 illustrates one embodiment of an electrical schematic of an
inverter
scheme that can be used to create AC power and control an embodiment of the
electromagnetic vibrator system. As shown in Figure 5, the inverter is
generally
comprised of a diode bridge and power transistors. In this embodiment,
insulated-
gate bipolar transistors (IGBT's) are used, though other power transistors as
known in
the art are contemplated with the scope of embodiments of the described
invention.
In one aspect, because of the large power requirements, the diodes and/or the
power
transistors may be operated in parallel. By operation of the power
transistors, the
control of the direction, and duration of current flow (and voltage polarity)
through
the load can be achieved.
[0021] The output of the inverter of Figure 5 can be used as an input to the
armature
winding of the transducer to control the operation of the transducer. For
example,
referring to Figure 6 it can be seen that a pulse-width modulated signal
(indicated by
8
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
an *) from an inverter such as the one shown in Figure 5 can be combined with
a
sawtooth wave (indicated by a + sign) to produce a desired sinusoidal wave
(indicated
by a ^). This sinusoidally-varying current can be used as an input to the
transducer
(armature winding), thus creating movement of the reaction mass of the
transducer
either up or down in a sinusoidal manner. Generally, the reaction of an
embodiment
of the electromagnetic transducer has higher fidelity to the input signal than
can be
achieved with a hydraulic shaker, thus fewer harmonics are imposed onto the
vibroseis analysis. Figures 7A and 7B illustrates simulated and actual voltage
and
current outputs from the embodiment of an inverter as shown in Figure 5.
[0022] In one embodiment as described herein, an electromagnetic vibrator
system
comprises an AC power source driven by an internal-combustion engine such as a
diesel engine; a power unit capable of providing the required electrical drive
to the
electromechanical transducer; an electromechanical transducer system that
converts
the electrical drive into mechanical vibrations to be transferred into the
ground; and
control and sensing circuitry, that insures that the energy transmitted into
the ground
faithfully represents the desired output. In one aspect, a described
embodiment of the
system can output typical linear and nonlinear vibroseis sweeps as well as
special
random sweeps developed for multi-source acquisition recording, and output a
sustained full ground force of 60,000 lbs.-force with a low end frequency
defined at
the -3dB point of full ground force of approximately 4 Hz and a high end
frequency
defined at the -3dB point of full ground force of at least 125 Hz. One
embodiment of
an electromagnetic vibrator system has the following parameters: low
distortion; 4Hz
to 225 Hz capability; 60,000 lb-force output from 4 to 125 Hz; force taper
from 125
Hz to 225 Hz; a reaction mass of approximately 16,500 lbs.; a base plate
having a
mass of approximately 4,707 lbs.; linear, non linear, and special random
vibroseis
sweeps; water (or other liquid) cooled for continuous operation; and a total
stroke of
approximately 6" to 6-3/4".
Examples
[0023] The following examples are put forth so as to provide those of ordinary
skill in
the art with a complete disclosure and description of how the compounds,
compositions, articles, devices and/or methods claimed herein are made and
evaluated, and are intended to be purely exemplary and are not intended to
limit the
scope of the methods and systems. Efforts have been made to ensure accuracy
with
9
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
respect to numbers (e.g., amounts, temperature, etc.), but some errors and
deviations
should be accounted for. Unless indicated otherwise, parts are parts by
weight,
temperature is in C, and pressure is at or near atmospheric.
Simulation Modeling
[0024] To provide a tool for evaluating the electrical and mechanical
performance of
the system design, a simulation code was employed. The code was implemented
using
the Simulink TM software, as known to one of ordinary skill in the art, which
includes electrical power block elements. Simulink TM employs a graphical
programming environment, so that the simulation layout is apparent visually,
and is
used to illustrate the discussion below.
[0025] The structure of the simulation model is shown in Figures 8A-8C. The
top
modular blocks consist of one block wherein run parameters can be set: sweep
time,
start and end frequency, component masses, ground model spring constant and
damping constant, etcetera. These constants are input to another block which
establishes the initial mechanical conditions for the reaction mass and base
plate.
[0026] Another section of the simulation (Figure 8B) uses initial conditions
along
with information about the actual computed motion of the reaction mass and
base
plate, and electrical information about the vibrator, to compute
electromagnetic force
being applied, and also the force produced by the ground on the baseplate.
[0027] The ground model consists of a spring and damper, which are coupled to
the
base plate, and the total hold down mass. The spring constant used for
simulations
was set to either model a hard clay surface, or alternatively a softer
response surface
such as sand. The magnetic force module incorporates both the current in the
force
coil and the position of the force coil relative to the base plate. In this
way the spatial
dependence of force generation, which was provided by finite element analysis
(FEA), was included in the model. In addition, a bi-directional spring model
was
employed to look for harmonic distortion effects. This model of the ground
used a
hard spring constant for the ground under compression, and a smaller constant
when
the ground was relieving from compression.
[0028] The third code section (Figure 8C) includes the vibrator control
module, the
AC generator prime power, a rectifier which takes the generator power and
charges a
DC link capacitor. For the simulation, the link voltage is 650 V.
[0029] Referring to Figure 8C, the controller sends a demand voltage signal to
a
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
PWM inverter which generates the demand voltage at the input of the vibrator
force
coil. The PWM module employs an algorithm which generates the true output
voltage
pulses without requiring use of actual switching thyristors. The PWM
architecture is
full H-bridge. Recent runs have used a 12 kHz switching frequency for the PWM.
[0030] The actual vibrator load was initially modeled as a simple inductor and
resistor. More recent runs have employed a more complex ladder network of
inductors and resistors, to capture the dynamics of the passive compensation
scheme
employed in the vibrator design. The electrical modeling is very detailed and
incorporates non-linear elements for force production, as well as PWM
switching
frequency effects.
A. Vibrator Controller Description
[0031] A block diagram of the control architecture is shown in Figure 8D. It
has a
module which lifts the reaction mass from an initial rest position to a height
at which
the force coil is centered in the DC magnet air gap. This is the start point
at which
vibratory sweeps begin.
[0032] The lift module uses an open loop algorithm to generate the lift to the
start
position. That is, a lift position versus time profile is used to generate a
force demand.
The profile used has no initial or final velocity. It has the form:
AX= i0 {1-cos((rt/Ot)} for t<Ot (1-1)
[0033] The lift module generates a current demand signal. Simulation results
show
that with this lift algorithm no start position over-shoot is produced. Once
lift is
completed, this module continues to generate a force demand which supplies
enough
force to compensate the gravitational force acting on the reaction mass. That
is, it
produces force to make the reaction mass neutrally buoyant.
[0034] The sweep control module accepts as input a file containing demand
force as a
function of time. To date, these files have consisted of linear sweeps, random
phase
sweeps, and random amplitude sweeps. The module employs feed forward and feed
back components to generate a current demand signal. A block diagram of the
sweep
control structure is shown in Figure 8E.
[0035] The demand force directly feeds into the feed forward module. The
difference
between demand and actual force generates an additional current command by use
of
a position-integral controller. The two commands are summed to generate a
total
11
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
current command.
[0036] It has been found that position control of the reaction mass can be
almost
completely achieved by selecting the initial start position of the reaction
mass and in
addition imparting an initial velocity to the reaction mass during the 0.1 s
before the
demand force is initiated. The initial velocity in particular has been found
to be very
effective in limiting the range of motion. This is supported by analytic
arguments. The
initial position and velocity are easily determined by double integration of
the demand
force table. This is done just prior to the sweep to set the appropriate
parameters in the
controller. The total current commanded, that is the sum of the commands from
the
lift and sweep modules, is input to a current control module. Its architecture
is similar
to that in Figure 8E. The command current generates a feed forward voltage
command, and also a feed back voltage command. The total voltage command is
output as the voltage demand signal sent to the PWM, which then generates the
input
voltage to the vibrator force coil.
B. Overall Performance
[0037] The simulation code was used to generate the electrical performance
characteristics of the major system components: the generator, the PWM, and
the
vibrator. A linear sweep, of 20 s duration, was employed starting at 2.5 Hz
and going
to 225 Hz. The ground force amplitude was 60,000 lb. Figure 8F shows
electrical
waveforms for the simulated vibrator.
[0038] The load power rises to nearly 300 kW at the end of the sweep. The peak
current level is less than 3.2 kA at all times, and the peak voltage 500 V.
The ground
force spectral power density is shown in Figure 8G.
[0039] The ground force increases with frequency, reflecting the gain
characteristics
of the force controller. This is correctable with modifications to the
controller.
[0040] The force control gain settings lead to more output than demanded at
higher
frequency. This causes the output power to be about 1 dB higher at 200 Hz. The
spectrum is fairly flat, and shows the power to be down less than 3 dB at 4
Hz.
[0041] The motion of the reaction mass for this sweep is shown in Figure 8H.
The
total range of motion is slightly less than 6 inches. The passive motion
control using
an initial position and velocity impulse keeps any late time drift of the
reaction mass
small.
C. Sweep Variation
12
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
[0042] An advantage of embodiments of an electromagnetic vibrator is its
ability to
track complex demand profiles with good fidelity. The case of a linear sweep
was
discussed above. More complex waveforms can also be utilized.
[0043] In one aspect, two different types of demand waveforms were simulated.
One
was a profile in which the wave phase was changed randomly in time (a random
phase sweep). The other was a profile in which the wave amplitude was changed
in a
random manner (a random amplitude sweep).
[0044] For the random phase sweep, the peak amplitude was held to about 60,000
lb.
Figures 81 and 8J show plots of demand ground force and actually produced
ground
force, respectively. Figures 8K and 8L show a comparison between spectral
power for
the two waveforms - demand and actual. A smaller frequency range is shown in
Figure 8L to indicate the good tracking of the vibrator output to demand
signal.
[0045] Figures 8M and 8N and Figures 80 and 8P show similar results for the
random amplitude case. Again the spectral power comparison is good.
[0046] To obtain information on how ground model non-linearity might generate
harmonic distortion in the vibrator ground force output, a bi-directional
spring model
was employed (described above). The variation in spring constant was set at
4:1. The
spring constants were also set to emulate sand conditions, i.e. a softer
spring constant.
Then sweeps at 40,000 pounds force were simulated at fixed frequencies. That
is, the
output was set to be monochromatic. Frequency values of 6, 25, and 125 Hz were
simulated.
[0047] Figures 8Q, 8R and 8S show spectral power plots for all three cases. No
harmonics are observed in the sweeps at 6 and 25 Hz. Very low amplitude 2d and
3Ta
harmonics were seen in the 125 Hertz case. The THD is essentially zero for all
runs.
Magnetic Analysis and Device Optimization
A. Geomagnetic Shaker
[0048] A geomagnetic shaker should have the ability to shake a large mass
through a
large stroke. Figure 9A shows an embodiment of a design suitable for this task
[2-4].
In one embodiment, there can be a DC power supply for the field coil and a
separate
PWM power supply for the armature coil that produces AC. In one aspect, a DC
winding in the armature winding produces lift, though in other aspects a DC
component from the PWM passes through all armature coils to produce lift. In
another embodiment, the PWM produces a DC offset of the AC armature current to
13
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
provide the lift. In one aspect, there is only one armature circuit and all
coils are in
series. The field coil carries dc current, and the armature carries both ac
and dc
current. The voltage demand on this coil can climb unacceptably unless one of
two
measures is adopted. The first is to place a passive copper (annotated copper
plating)
band on either side of the armature coil. The induced currents are essentially
opposite
in phase to those in the armature. The second option is to wind a second
winding on
the central and outer pole which is connected in series with the armature
winding.
This so-called active compensator is effective only when it shadows the
armature coil.
Thus it has the two negative features of not being able to compensate when
longer
strokes are involved and the added construction difficulty of running armature
current
to the same structure that the field coil is mounted to. There is actually a
more
fundamental aspect of this trade in considering whether to use a passive
versus an
active compensator, having to do with a theoretical limit.
B. Passive Compensation
[0049] Passive compensators can have a lower inductance at higher frequency
than
active compensation. Figure 9B shows a thought experiment that can be
performed on
two coils to defend this statement.
[0050] To simplify this analysis, consider the case where LI = L2 and RI = R2.
At
radian frequency cc, the voltage across the two coils in Figure 9B(a) is
[0051] V =2=(jcoL, +R,)i, +2jcoMi,. (2-1)
[0052] Let M = k L,L2 . The inductance for case (a) is
[0053] Leq = s V = 2L, (l+k). (2-2)
[0054] The shorted coil of case 9B(b) requires a prediction of current. Since
it is
shorted,
[0055] (jwL2 +R2 )i2 + jwMi, = 0. (2-3)
[0056] Solving for i2,
[0057] i2 jwMi, . (2-4)
jwL2 + R2
[0058] Inserting (2-4) into the voltage equation for Vl gives
= jwL,i, + 2M2i, . (2-5)
[0059] V
jwL2 + R2
14
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
[0060] Writing (2-5) in terms of real and imaginary parts gives
[0061] V =JwiI L, - 0) 2M2L2 + R2w2M2i, +R1i1. (2-6)
P22 + w2L 2 R22 + 0)2L22
2 [0062] The first term in parenthesis on the right-hand side of (2-6) is the
equivalent
inductance of this circuit. At high frequency co >> [3J.
L2
[0063] Leq = L1 (1- k 2 = (2-7)
[0064] Consider a coupling coefficient with k = -0.95. The anti-series
solution yields
Leq = 0.1 L1. The passive shorted coil yields Leg = 0.0975 L1. In the high
frequency
limit, the passive compensator always requires less voltage since it has the
lower
inductance. The two are equal when k = 1. Note that (1-k2) = 2(1+k) when k = -
1. For
coupling coefficients -1 < k < 0, the passive compensator always has the lower
inductance.
C. The Role of the Compensator
[0065] Consider the geometry shown in Figure 9C. What is the effect of the
compensation coil on force if the center two excitation coils are allowed to
move
vertically up the open channel, along the highlighted segment? Do this with
the
compensated coils activated so that they carry exactly the same current in the
opposite
direction. Repeat the experiment when the compensation coils are unexcited
(removed). Figure 9D shows that the uncompensated case has the flatter force
curve
versus position. Next, reverse the current in the excitation and compensation
coils so
that the force is downward. Figure 9E and Figure 9F show that the compensation
winding reverses the slope of the force versus position curve. An important
point is
that the mean of the compensated and uncompensated forces are identical to
within
three decimal places.
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
D. Design Geometry
[0066] The primary difference between a geomagnetic seismic shaker (Figure 9G)
and an electromagnetic shaker affixed to a shaker table is that the field coil
moves in
the seismic shaker. An example of design criterions are as follows: ac force
generated
= 267 kN (60 klb); dc force generated = 75.6 kN (17 klb); total force
generated on the
down stroke for the armature = 342.5 kN (77 klb); Weight < 7.73 kg (17 klb);
Stroke
allowance 17.14 cm (6.75 in.); and Power < 448 kW (600 kW).
[0067] The design variables as shown in Figure 9H should be chosen carefully
to
achieve these objectives. In addition to the variables shown, the force under
maximum
excitation must be computed in various positions y to ensure that the force
objectives
are met as an average over the entire displacement range. The down stroke
force
condition dominates in this design. The down stroke condition is accompanied
by a
sizeable reluctance force to pull the armature up further in the slot. This
force is
nearly constant and close to 177 kN (40 klb) to pull the armature into the
field coil
regardless of the sign of the armature current.
[0068] The optimization is approached using a trust region optimization
algorithm.
Sequential quadratic programming (SQP) algorithms are among the more powerful
deterministic methods that use gradient information [2-5, 2-6]. The response
surface
method attempts to build an n dimensional surface of the problem and then
examines
how that surface varies with the system unknowns [2-7]. Something akin to the
response surface method in [2-8] is used to determine the right shape,
minimizing the
power dissipation subject to the constraints listed previously.
[0069] Figure 91 shows a three-dimensional surface linking power to two of the
variables. In reality a five-dimensional surface is constructed in this
optimization for
each of the key variables, source power, weight, and force. This five-
dimensional
spline fit represents the data; it can also be used to compute the derivatives
for all the
variables [2-9]. Knowing the gradient information allows for a fast and
accurate
determination of minimum power satisfying the constraints [2-10,2-11,2-12].
[0070] Each of the five variables annotated in Figure 9H is varied in a nested
loop.
Force has to be computed at a minimum of four positions. As far as the finite
element
analysis is concerned, these four positions serve as a sixth variable. Suppose
that the
variables are perturbed about four positions. The number of finite element
analyses
required would be:
16
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
# analyses = 46 = 4096. (2-8)
[0071] Only negative forces (forces to increase the separation of the armature
and
field winding) are examined at the top of the stroke and only positive forces
at the
bottom of the stroke. This is close to what is observed in simulation.
[0072] Current density is central to the solution of the nested set of
problems. This
density is smeared over the current carrying regions shown in Figure 9J.
E. Current Density
[0073] Onsite testing has shown that a 3.175 mm (0.125 in.) diameter copper
tube
with a 1.58 mm (0.063 in.) diameter water cooling hole can sustain 200 A
steady
state. This corresponds to a current density of
200.4
J - =39.372 =3.368.107A/m2.
~` 7c(1/64-1/162)
(2-9)
[0074] The conductors for the field and the armature are targeted at 12.7 mm
(0.5 in.)
square with a 6.35 mm (0.5 in.) cooling hole. The area open to copper is
A.=0.52-TC )2 / 4 = 0.201 in 2.
C
[0075] 4 (2-10)
[0076] For the field windings they can be insulated with a 10 mil wrap of
glass
followed by a 2 mil wrap of KaptonTm. So the equivalent current density in the
field
winding is
0.524 2 = 2.464 _107 A /M2. (2-11)
[0077] J field = J. A
[0078] The armature current density must account for the tie dependence of the
waveform. A steady state base force fb equal to the weight of the unit (75.6
kN (17
klb)) must be superimposed on an oscillating component capable of achieving a
target
force f (267 kN (60 klb)). The current density in Eq. (2-9) is a steady state
value.
What current density in the armature J,,,,n will deliver the same heating in
the copper?
That value is determined by solving the equation
z
[0079] J 2At = r[j fib )+( f` sin (co t) dt. (2-12)
Jb+f fib+f
[0080] Average loss is of interest. The integration needs to be performed over
an
integral number of periods. When this is done, the allowed current density
that exactly
equals the loss of a dc coil carrying JJõ is determined
17
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
_ Jcu _ 'Jcu
'farm 2 _.' z - ,
{ .fb + f /3z + 2 (1 _ fl)z
Jb+f 2 fb+f
[0081] where (2-13)
fb
f + fb
[0082] The multiplier on Jeõ is an interesting number. When the base weight
(force) is
a large proportion of the total force, this multiplier goes to unity. When it
is a small
proportion of the total weight, i.e., the desired force is nearly sinusoidal,
this
multiplier goes to h . Figure 9K shows how this multiplier changes beginning
with a
100% sinusoidal load to a purely do load at /3=1. The interesting point is the
maximum value (,13-=1.73). The large multiplier results due to the combined
sinusoidal effect with an offset. This density must be reduced by the packing
density
as in Eq. (2-11) to get an equivalent modeling density. Since 20 mil
insulation wrap
followed by a 2.5 mil KaptonTm is envisioned, the equivalent peak current
density
should be
[0083] J = J. = 3.275.107A/M2. (2-14)
J. - 0.59Jfl2+1(1fl)2 F. Power Calculations
[0084] The current densities listed in Eq. (2-11) and Eq. (2-14) are apparent
current densities that account for fields and forces. The common expression
for
computing power is
p = (JA)2 t = J2 Vol. (2-15)
CA a
[0085] The apparent volume Vap used in a field calculation is larger than the
actual
copper volume V. This enters as a correction which requires increasing the
apparent
power demand. For the field, the power dissipated in the field is
z
Pfeld - Jfeld 2 Vap = J1õ 2 VaP . V. (2-14)
6Vau 6 VaP
[0086] where Jf:eld is defined in Eg. (2-11). A similar correction must be
applied on
the armature power. It is easier to use J,a in this calculation since the
equivalent Jar',
was defined to match it.
18
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
P. = J 2 Va = V. . (2-17)
a. r 6 Vap
[0087] Here Vap is the apparent armature copper volume. This is of course the
loss
ignoring skin proximity effect. These effects are considered after the
optimization
since it is simply not practical to add this complexity to the optimization
procedure.
G. Optimization Results
[0088] Figure 9L shows an embodiment of an optimal design minimizing power
dissipation within the constraints. Homogeneity in the field plot shows that
the steel is
used nearly to the maximum throughout Figure 9M.
[0089] The optimized armature cutout closely fits a series of 6.35 mm (0.25
in.)
conductors as shown in Figure 9N. The voltage limitations of the inverter
electronics
suit two parallel circuits more than one series circuit. The staggering
registered in this
layout is an attempt to balance the induced voltages.
[0090] There are two ways to introduce the dc component of current required in
the
armature to achieve static lift. The first is to ask the inverter to supply
that added do
component. The second is to actually excite a subset of the inverter coils to
carry only
dc. Since do power supplies are cheaper than ac inverters, this second option
saves
money.
[0091] Assuming this option is adopted, the question as to where to place the
dc
armature coils surfaces. The ac excited armature coils induce parasitic losses
in the do
sub-conductors. The simulation summarized in Figure 90 shows that a power
saving
of 10% is realized if the dc coils are positioned close to the center of the
armature
stack.
[0092] The optimization procedure computed a desired field excitation of 134
kAT.
Total field power loss in all field coils plus eight armature coils containing
dc is 275
M. The real ac armature power loss at 225 Hz is approximately 306 M. The
magnetic field generated by the field coil alone is shown in Figure 9P.
H. Power Dissipation with Skin and Proximity Effect
[0093] The only effective way to achieve the forces demanded with the weight
constraints is to use the large current densities as described herein. These
internal
water cooled coils suffer from parasitic skin effect and proximity effect
losses as the
armature frequency increases. Skin effect, moving the current to the outside
of the
conductor is well understood.
19
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
[0094] In each of the plots shown in Figure 9Q, a fixed current was specified,
but not
the distribution. For this axi-symmetric problem, the field can be
characterized by a
single component of magnetic vector potential A0. A finite element code was
used to
solve
OZA0 - jaqu6AA = pJ5. (2-18)
[0095] Here .15 is the source current, except in this current imposed
conductor where it
is assigned to be zero. In the conductor the vector potential is assigned to
have two
parts, a constant component A, and a variable component to be determined. If
the
current in the sub-conductor is I, A, has the value II(j&So)), where a is the
conductivity of the region and S is the regions surface area.
[0096] As expected, the low frequency current density distribution shown in
Figure
9Q(a) is essentially uniform. Figure 9Q(b) shows the distribution of current
in several
sub-conductors that are clustered at 225 Hz. Figure 9Q(c) shows the
distribution of
current in the same sub-conductors adjacent to a copper compensation slab. The
current induced in the slab is 180 degrees out of phase with the armature
conductor
current. The maxim about unlike currents attracting applies, and explains why
the red
currents crowd towards the center.
1. Power Calculations and Equivalent Circuits
[0097] Minimizing power is of preeminent importance since the generator must
be
carried and supplied by the vehicle carrying the equipment. One way to compute
the
inverter power is to shut down the dc field coils, excite the armature only
with ac, and
integrate t x H over the interface of all regions carrying current. It should
be clear
that the presence of a compensator can lower the voltage required on the
armature
coil. As already shown in Figure 9H, the presence of compensation does not
alter the
average force, only its mean. Since inverters are rated by their volt-amp
product, this
is of no small importance. The reader should note that the current induced in
a passive
compensator will never exceed that in the armature circuit. Thus a thicker
compensator should also result in a smaller real power requirement.
[0098] The tabulated results annotated in Figure 9R show that both voltage and
power
requirements are reduced with the thicker 12.7 mm (0.5 in.) compensator. The
voltages are computed by dividing the volt-amp product by the current.
J. Control
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
[0099] In addition to a stabilized lift generation, it is necessary to
generate a number
of force profiles including those that linearly increase with frequency. The
controller
requires a good circuit model in order to supply the right control voltage.
The
presence of the compensator complicates the use of an equivalent circuit. Both
Silvester [2-13] and Giesselmann [2-14] approach this type of diffusion system
with a
ladder circuit such as that shown in Figure 9S. The BLv voltage is a motion-
induced
voltage. This term is computed by computing the flux 0 linking the armature
coil as a
function of position with the dc field excitation on. The BLv voltage for any
velocity
is
BLv = d v. (2-19)
y
[00100] The derivative is determined by fitting the flux to a spline and
taking the
derivative of the spline [2-9]. The ladder parameter fit was determined using
a
Levenberg-Marquardt method with line-search to best fit the real and imaginary
power dissipation [2-15]. Figure 9T shows that this representation method is
quite
good for a large range of frequencies.
K. Closing the Design Loop
[00101] Closing the design loop is equivalent to asking whether the device
meets the
force requirement with rated current. The highest frequency is the most
taxing. What
is the force when the field coil is excited with dc, the armature coil with
ac, and the
eight subconductors in Figure 9N excited with dc? The analysis must be a
transient
time-stepping analysis. Eddy currents in both the steel and the compensator
will only
respond to the ac current. Figure 9U shows that the lift force (negative) does
indeed
reach the target of 77 kib in the presence of all compensation and eddy
currents.
L. Permanent Magnets - A Possible Way Ahead
[00102] The power demand for the device may be a concern. Of the 275 KW do
loss,
260 kW comes from the field winding alone. Options for reducing the power
dissipation include using smaller conductors for the armature winding to
reduce skin
and proximity effect and magnets to replace the field winding. Magnets can
have the
greatest effect on real power usage. Figure 9V shows two columns of 12.7 mm
(0.5
in.) 45 MGO NdFeB magnets placed on either side of the armature gap in the
absence
of any field coils. Forces are within 7% of the target result with this
arrangement.
Since the magnets cannot be turned off, a concern about interference with
tools might
21
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
result from their use. The leakage field from these magnets is shown in Figure
9W. A
leakage field of 0.5 T can be expected 7.6 cm (3 in.) away from the armature
gap.
M. Conclusions
[00103] A magnetic optimization has been performed to design an
electromagnetic
shaker capable of generating 342 kN. The response method approach employed in
this
exemplary optimization required 4094 finite element analyses per optimization
round.
The design can accommodate a stroke length of 17.1 cm (6.75 in.) and the total
armature and field base weight is just under 7.72 kg (17 klb). A transient
analysis
shows that the design meets the force specifications. Permanent magnets appear
to be
an option for reducing real power consumption.
N. References
[00104] Unless otherwise noted, each of the references listed below are fully
incorporated by reference into this paper and made a part hereof-
[2-1] H. Dai, M. Sain, and B. Spencer, "Using tensors to track earthquakes on
hydraulic
shaker tables," IEEE Trans. on Mechatronics, vol. 4, no. 1, March 1999, pp. 60-
70.
[2-2] L. Flora and H. Grundling, "Acceleration control of an inverter-fed
electrodynamic
shaker," IEEE Power Electronic Specialist Conference, June 18-22, 2006, pp. 1-
7.
[2-3] R. Fair and H.R. Bolton, "Analysis and design of electromagnetic moving
coil
vibration generators,"1993. Sixth International Conference on Electrical
Machines
and Drives, Sept. 8-10, 1993, pp. 529-534.
[2-4] M.-T. Peng and T. J. Flack, "Numerical analysis of the coupled circuit
and cooling
holes for an electromagnetic shaker," IEEE Trans. on Magn., vol. 41, no. 1,
Jan. 2005,
pp. 47-54.
[2-5] R. Fletcher and M.J.D. Powell, "A rapidly convergent descent method for
minimization," Computer Journal, vol. 6, 1963, pp. 163-168.
[2-6] Goldfarb, D., "A family of variable metric updates derived by
variational means,"
Mathematics of Computing, vol. 24, 1970, pp. 23-26.
[2-7] G.E.P. Box and N.R. Draper, Empirical model building and response
surface, John
Wiley and Sons, New York, 1987.
[2-8] P. Alotto, P. Girdinio, P. Molfino, and M Nervi, "Mesh adaption and
optimization
techniques in magnet design," IEEE Trans. on Magn., vol. 32, no. 4, July 1996,
pp.
2954-2957.
[2-9] Carl de Boor, A Practical Guide to Splines, ISBN 9780387953663, Applied
22
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
Mathematical Sciences, New York, Springer, 1978.
[2-10] K. R. Davey, "Magnet design optimization using variable metrics," IEEE
Trans. on
Magn., vol. 31, no. 6, 1995, pp. 3566-3568.
[2-11] K. R. Davey, "Use of tensor product splines in magnet optimization,"
IEEE Trans. on
Magn., vol. 35, no. 3, May 1999, pp. 1714-1717.
[2-12] K. R. Davey, "Examination of various techniques for the acceleration of
multivariable
optimization techniques," IEEE Trans. on Magn., vol. 39, no. 3, May 2003, pp.
1293-
1296.
[2-13] P. Silvester, "Modal network theory of skin effect n flat conductors,"
Proc. IEEE, vol.
54, no. 9, Sept. 1966, pp. 1147-1151.
[2-14] M. Giesselmann, T. Heeren, A. Neuber, and M. Kristiansen, "Advanced
modeling of
an exploding flux compression generator using lumped element models of
magnetic
diffusion," IEEE Pulsed Power Plasma Science Conference, June 17-21,2001, vol.
1,
pp. 162-165.
[2-15] D. Marquardt, "An algorithm for least-squares estimation of nonlinear
parameters,"
SIAM Journal Applied Math., vol. 11, 1963, pp. 431-441.
Insulation Design
[00105] The class rating for insulation systems indicates a temperature range
where the
resin in the insulation starts to sublime and over time can lose insulation
strength.
This rating is appropriate if the conductors are not carrying much mechanical
load
which is the case for winding in iron slots of generator stators. The coils
for the
electromagnetic vibrator (EMV) have mechanical load and the rating that must
be
observed is the Heat Deflection Temperature (HDT). Figure 10A shows laboratory
test data from which the HDT and Glass Transition Temperature (Tg) are
calculated.
The figure also presents a table of these values for some of the common resin
systems.
The HDT can be thought of as the maximum temperature at which the insulation
retains good strength and Tg is the temperature at which the insulation will
show
duress under load.
[00106] The layup of insulation on the conductor is shown in Figure 10B. The
KaptonTm can be bonded to the copper tubing with Teflon through a heating
process. This produces a very robust insulator. Two more layer of KaptonTm can
be
applied to the conductor and then an overwrap of glass can be applied as the
armor.
The glass readily wets with the VPI resin and creates a very strong mechanical
23
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
encapsulation of the conductor. The figure also indicates the insulation
strength for
this system. It is an order of magnitude greater than the highpot inspection
voltage.
This specification should provide high temperature capability and long life.
EMV Mechanical Design
[00107] The major mechanical parts for an embodiment of the EMV comprises the:
1) Reaction Mass
2) Stainless Steel Base Plate
3) Field Coils
4) Force Coils
5) Passive Compensation Copper Conductors
6) Guide Bearings for Base Plate
7) Robust Electrical and Cooling Terminations
Many of these parts are seen in the EMV baseline design Figure 11A.
A. Reaction Mass
[00108] In one embodiment, the reaction mass is a three part assembly made of
a type
of steel whose alloy properties are most favorable to conducting the magnetic
fields
generated by the EMV field coils. In one aspect, this low carbon steel is
commonly
known as AISI 1005, but is referred to as A216, if discussed in context to a
casting.
The reaction mass (in position over the base plate) is shown in Figure 11B.
B. Stainless Steel Base Plate
[00109] In one embodiment, the base plate is a large casting of non-magnetic
CF8
(304) corrosion-resistant material as shown in Figure 11C.
C. Field Coils
[00110] In one embodiment, the field coils in the baseline design are
constructed using
.25in. square copper conductor tubing with a .125in diameter center hole. A
mixture
of water and ethylene glycol can flow through the coils to remove the heat
generated
from the current flow through the coils when the EMV is running. The field
coils can
be contained in a machined pocket in the top piece of the reaction mass. The
coils can
have several layers of insulation. The outer layer can be 5mil glass cloth
that can be
vacuum impregnated (VPI) with a high temperature rated (class H) resin. The
field
coils and their location in the reaction mass can be as shown in Figure 11D.
D. Force Coils
[00111] In one embodiment, the force coils are also be wound out of 0.25 in.
square
24
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
copper conductor tubing with a 0.125 in. diameter center hole. These coils can
also
be actively cooled. They can be secured to the base plate as shown in Figure 1
lE and
can also be resin impregnated like the field coils.
E. Other EMV Components
[00112] Other important parts to the EMV include the passive compensation
copper
rings, the guide bearings and robust electrical and cooling terminations.
These
component parts are seen in Figure 11F. Many of these can be custom designed
and
fabricated. One embodiment of the fabrication steps are shown in Figures 11 G
through 11-J, resulting in the final assembly shown in Figures 11K and 11L.
Vibrator Power Supplies
[00113] The power electronics section comprises several components of an
embodiment of a system, starting at the inverter that powers the force coils.
This is an
important power component as it powers the force coils with the current that
is fully
controlled by the system controller. The inverter has a DC input from the link
capacitor and an AC output to the force coils. The output frequency is
required to
sweep from 4 to 225 Hertz while maintaining a sinusoidal output with very low
THD,
Total Harmonic Distortion.
[00114] In addition to the inverter output function, the inverter package can
also
include a three phase rectifier bridge and the link capacitors. The rectifier
bridge is a
three phase, full-wave, design. The upper set of switching devices can be
controlled
devices while the lower three can be passive rectifiers. Making this a half-
controlled
bridge can allow for controlled soft-start charging of the link capacitor from
the
alternator. After the initial charging is completed, the unit can function as
a
traditional full-wave bridge.
[00115] In one embodiment, the DC link capacitor value was selected to be 0.1
Farads.
The link capacitor can provide energy for the inverter in the quick burst that
the
inverter demands. The generator is the source of the energy, but it can
operate at 70
Hz and may not be able to directly supply the inverter. The link capacitor
needs to
have enough energy available to assist during the low frequency portion of the
sweep
which is what dictated the 0.1 F value. The capacitor voltage is set by the
generator
output voltage, 480 V rms in this case, to be 680 V dc.
[00116] The next component in the line is the inverter itself. The inverter
can be a
single phase unit, configured in a classic `H' bridge layout. This unit may be
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
mounted on a truck and operate in all possible conditions, therefore the
ambient
temperature can vary from -35 to 60 C. The expected coolant inlet temperature
is no
higher than 65 C. The unit can be housed within a NEMA 4 style enclosure. The
required output current is 2400 A rms at up to 350 V rms. The duty cycle is
100%.
To achieve a low THD value and provide a swept frequency output of 4 to 225
Hz, it
is estimated that a minimum switching frequency of 2.5 kHz is required.
[00117] In one embodiment, two separate inverter units can be connected in
parallel to
meet the current (e.g., 2400 A) specification. Each unit can be comprises of a
three
phase unit collapsed down to a single phase output. Such an inverter is
available
from, for example, Applied Power Systems, Inc. of Hicksville, NY.
[00118] In another embodiment, the inverter can be a highly parallel design
switching
at 2.5 kHz with a dead-time minimum value of 3.2 microseconds. This inverter
is
using four Eupec 1200 A IGBTs in parallel to achieve an output current rating
of
2400 A rms in the low frequency range of 20 Hz and lower. In the upper range
of 20
to 250 Hz the output rating is 3000 A rms. The expected distortion is 2.5 %
THD.
Such an inverter is available from, for example, Saminco, Inc. of Fort Myers,
Florida.
[00119] In one aspect, a DC power supply is for powering the field winding. In
one
embodiment, the requirements are for 931 A in the field coil. To meet this
specification, an adjustable power supply with an output of 0 to 300 Volts and
0 to
1000 A can be used. Such a DC power supply is available from, for example,
Dynapower Corporation of South Burlington, Vermont.
[00120] In one aspect. power for the inverter and the dc power supply can be
provided
by an alternator connected to a prime mover such as a diesel engine. The
alternator
specifications are based on the loads it will drive. In one aspect, the
alternator can be
a three phase, 480 Vac, 570 kW unit having a 60 C ambient temperature rating.
Such
an alternator is available from, for example, Marathon Electric of Wausau, WI.
In
other embodiments the architecture of the DC power supply can be integrated
with the
PWM to provide both functions from one highly integrated cabinet.
Controls
[00121] Figure 12A is a block diagram of the components of an exemplary
control
system. From the top level down, the control system includes the system
controller
board, operator panel, down to an array of sensors. In one exemplary aspect, a
dSPACE control board as available from dSPACE, Inc. of Wixom, MI can be used
to
26
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
allow for integration of the Matlab or Simulink test environment. In a
prototype test
environment, the flexibility of using a simulation package such as Simulink
for the
controls algorithm is ease of code changes. The dSPACE board is powered by
PowerPC 750GX processor with a 1 GHz clock speed. The board has 20 analog to
digital, A/D, channels with a +/- 10 Volt range. These A/D channels are all 16
bit
vertical resolution. There are 8 16 bit D/A channels, also with a +/- 10 Volt
range.
There are 32 digital channels that can be configured as input or output
channels. The
dSPACE board includes a mounting enclosure, and a breakout panel that allows
for
easy connection to the input and outputs.
[00122] The block diagram (Figure 12A) shows the operator panel in the lower
left
area. This can be a panel with the basic controls for operating the EMV. The
panel
can allow for manual operation and testing of the various system components.
During
normal planned operation, all of the switching can be done using the dSPACE
board
in an automated fashion. The operator panel can have indicators for the status
of
many of the major components. This can assist in completing the initial
checkout and
locating any problem areas.
[00123] The controller board can handle all the system operation functions.
The
controller can collect and store signal data for later analysis. These
include, for
example: PWM control and gating; DC power supply control; Alternator
functions;
Accelerometers; Position sensors; Current sensors; Voltage sensors;
Temperature
sensors; and Coolant pumps.
[00124] The control board can be able to play the rising frequency sweeps, or
take in
external data files for an input for random phase sweeps. Again, the most
flexibility
possible can be built into the system to allow for the widest possible range
of tests.
Exemplary Truck Layout
[00125] Described herein are basic modifications to a hydraulic shaker truck
(i.e., the
Mertz truck) to allow the installation of an embodiment of an EMV as described
herein.
[00126] As shown in Figure 13A, the generator can fit directly behind a diesel
engine
and the hydraulic drive transmission. The generator can be directly driven
from the
take-off pad that did drive the old vibrator hydraulic drive pump. The removal
of the
un-used vibrator hydraulic drive pump frees up the center most drive position.
An
adapter flange can connect to the SAE pad, and a spline shaft can couple to
the drive
27
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
pad. This adapter can drive a short driveshaft that can connect to the
generator shaft.
This setup can avoid requiring any sort of gearbox. The hydraulic drive does
not
include any sort of gear change. This exemplary arrangement requires the
generator
to operate at the Detroit diesel speed of 2100 RPM, instead of the generator
design
speed of 1800 RPM. The increased generator speed can result in an increase of
the
output frequency to about 70 Hertz, instead of 60 Hertz. The Marathon
generator that
has been selected has a maximum speed of 2250 RPM. To achieve an acceptable
level of angle on the short driveshaft, the diesel engine should be relocated.
The
engine is mounted on a skid which is attached to the Mertz frame. The skid
also
includes the radiator, the hydraulic drive transmission, and many engine
accessories.
This can make relocating the engine rearward about 20" and upward 8" a simple
task.
A new U shaped skid can be made out of three pieces of channel steel, then
slipped
under the existing skid. The Mertz frame extends back the extra length,
however the
spare tire may be mounted elsewhere. Two new brackets that secure the skid to
the
frame can be fabricated. Some of the existing hydraulic hoses may not be long
enough, which may require some adapters and new short lengths of hose to
extend
them to the rear.
[00127] Other modifications may include moving the diesel fuel tank from
behind the
cab to the center of the truck over the quad. The existing hydraulic oil tank
can be
rotated 90 degrees from next to the quad, to the right side of the truck.
[00128] Besides the generator, other major components that can be added to the
truck.
These are the inverter, the DC power supply, and a heat exchanger for the
force coils.
The space behind the cab can hold the inverter and the DC power supply. These
are
shown in the truck layout Figures 13A, 13B and 13C. The heat exchanger can be
added on the passenger side of the truck, directly behind the existing heat
exchanger.
The heat exchanger fans can be driven in the same manner as the existing fans,
with
small hydraulic motors powered off the same source.
Baseplate Design and Analysis
A. Introduction
[00129] Structural analysis of the baseplate design was done in order to
determine the
maximum operating stress and whether or not the baseplate remains in contact
with
the ground under dynamic loading conditions. In addition to the dynamic
analyses, a
static analysis was done to determine stress due to off-axis loading of the
baseplate
28
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
that would be encountered, for example, if the baseplate was operated on a
hill-side.
The goal of the analysis was to determine the optimal baseplate configuration
that
would minimize weight while retaining stiffness and strength. The dynamic
analyses
were done using ABAQUS TM, a commercially available finite element software
package with full transient and nonlinear analysis capabilities. The static
analysis was
done using COSMOSWORKS TM, a finite element analysis supplement to
SOLIDWORKS TM solid modeling software.
B. Dynamic Response Models
[00130] Transient response of the baseplate to sinusoidal excitation of the
reaction
mass at three distinct frequencies was done using an explicit time integration
scheme
available within ABAQUS TM. The explicit integrator was chosen over the
implicit
integrator because of superior performance in models that involve contact.
Even
though the explicit scheme requires very small time steps (on the order of 10-
7 sec)
due to its conditionally stable nature, run times for large models are
generally shorter
than for an implicit scheme since the stiffness matrix does not need to be
decomposed
and no intermediate iterations are needed for convergence in each time step.
Furthermore, the small time step required ensures that the response is
captured
adequately. With implicit schemes, attempts to shorten run times by increasing
the
time step size could result in undesired loss of solution accuracy.
[00131] Figures 14A and 14B show the finite element model used in the dynamic
analyses. Figure 14A shows the extent of the ground that was modeled along
with the
baseplate, and Figure 14B shows the baseplate only. Symmetry of the baseplate
and
loading conditions allowed reduction of the problem to a quarter-symmetry
model,
thus allowing a higher mesh density for a given number of degrees of freedom.
Due
to the complex shape of the baseplate structure, tetrahedral elements were
used to
mesh the solids using the automatic mesh generator available in ABAQUS TM.
Second order interpolation functions were prescribed in order to maximize
resolution
of stress gradients within the structure. The mesh density chosen was a
compromise
between solution convergence and turn-around time, considering the limited
time
available for parameter studies in the initial design phase. Future work can
address
the issue of whether convergence was achieved at stress concentrations.
C. Dynamic Load Cases
[00132] Analyses were done to simulate operation of the vibrator at 10 Hz, 125
Hz and
29
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
200 Hz. These three frequencies were chosen to provide snapshots at the low,
mid
and high frequency ranges of operation. Since the analysis turn around time is
significant in a transient analysis, it was not practical given the time
constraints to
model a swept sine input that would involve several seconds of load duration
and
several million time steps. For the 10 Hz case, the analysis was run out for 3
periods
of load fluctuation. For the 125 Hz and 200 Hz cases, the analyses were run
out for 5
periods.
[00133] In all three cases, a 53,000 lb distributed load was applied around
the
perimeter of the baseplate to simulate support of the vehicle (less the
baseplate and
reaction mass) on airbags. A 17,000 lb load was applied to the upper surface
of the
cylindrical portion of the baseplate to simulate the steady state (also
referred to as the
D.C.) load of the reaction mass acting through electromagnetic forces on the
baseplate
coils. Finally, a 60,000 lb alternating load was applied at the same surface
to simulate
the reaction mass being driven electromagnetically. The alternating load was
varied
sinusoidally at 10 Hz, 125 Hz and 200 Hz in three separate analyses.
D. Ground Model
[00134] The ground plays a part in the dynamic response of the baseplate, and
it was
included in the model. The ground was modeled as a hemisphere with a 240 in.
radius. Of course, due to symmetry, one quarter of the hemisphere was actually
modeled. An initial analysis was done with a soil modulus of 200,000 psi and a
density of 0.052 in/lb3. Per recommendation by Ion, the soil modulus and
density
were adjusted to produce a 30 Hz frequency by lowering the modulus to 20,000
psi,
and increasing the density until a model of the ground without baseplate
resulted in a
fundamental natural frequency of 30 Hz. These values of soil modulus and
density
were used in subsequent analyses.
E. Boundary Conditions and Contact Surface Definition
[00135] Symmetry boundary conditions were applied at model faces corresponding
to
the cutting planes for the quarter-symmetry model. In addition, the outer
spherical
surface of the ground model was fixed in all three directions. A contact
surface was
defined between the ground and the baseplate that allowed separation of the
baseplate
away from the ground but no penetration.
F. Static Analysis Model
[00136] Figures 14C and 14D show the finite element model used for the static
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
analysis. In this case, the lateral loading condition allows a half-symmetry
model but
not a quarter-symmetry model. Since the analysis run time is orders of
magnitude
smaller than that needed for the dynamic analyses, the mesh density was
increased to
provide a reasonably fine mesh at stress concentrations. Initial runs
indicated the
highest stress was in the region of a vertical slot in the upper cylindrical
portion of the
baseplate, and the mesh was refined in that region accordingly, as seen in
figure 14D.
G. Static Load Cases
[00137] For the static analysis, the 60,000 lb alternating load was applied as
a static
load acting along the axis of the vibrator. The 53,000 lb vehicle weight load
was
applied around the perimeter of the baseplate, and the 17,000 lb reaction mass
D.C.
load was applied at the upper surface of the cylindrical portion of the
baseplate. The
17,000 lb reaction mass load was decomposed into a vertical and a lateral load
to
model a 15 degree tilt of the baseplate on the ground. Thus the 17,000 lb
reaction
mass load was modeled as a 16,420 lb vertical load and a 4400 lb lateral load.
H. Boundary Conditions
[00138] The ground was not modeled in the static load case, thus the bottom of
the
baseplate was fixed in all three direction to prevent rigid body motion.
Symmetry
boundary conditions were applied at the half-symmetry cutting planes.
1. Results
[00139] Initial analysis indicated that the inner portion of the baseplate
separates from
the ground whenever the sinusoidal load periodically acts in an upward
direction. The
outer perimeter, however, remains in contact with the ground at all times due
to
application of the truck weight through the airbags located around the
baseplate
perimeter. In an attempt to eliminate the separation, the baseplate thickness
was
increased from 2 to 3 in. The thicker baseplate did reduce the amount of
separation
from the ground, but did not eliminate it entirely. The analyses presented
here are for
the thickened baseplate configuration.
[00140] Table 1, below, presents the maximum Von Mises stress and the maximum
gap between the ground and baseplate encountered in the three cases. Stresses
for all
three cases are quite low compared to material strength. Deflection of the
baseplate
away from the ground is higher at low frequency, suggesting that the ground
resonant
frequency is coming into play.
[00141] Figure 14E shows an exaggerated deflection plot of the baseplate
lifting up
31
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
from the ground during the course of the analysis. As seen in the plot, the
outer
perimeter of the baseplate stays in contact with the ground and only the
central
portion lifts up off the ground.
[00142] Figures 14F and 14G show Von Mises stresses in the baseplate. Figure
14F
shows the overall maximum Von Mises stress encountered anywhere in the
baseplate
in the three loading cases, and Figure 14G shows the maximum Von Mises stress
encountered in the cylindrical portion of the baseplate. Both stresses are low
compared to the material strength.
[00143] Figure 14H shows the maximum Von Mises stress from the static
analysis.
The stress is 3.2 ksi, and is also quite low compared to the material
strength. Figure
141 shows the extent of lateral deflection induced by the lateral load. The
deflection
is less than 0.0012 in.
Table 1. Results from dynamic analyses
Load case Max. separation from ground Max. Von Mises stress
Hz 0.0175 in. 6.4 ksi
125 Hz 0.009 in. 7.5 ksi
200 Hz 0.010 in. 14.3 ksi
J. Summary of Structural Analyses
[00144] Stress results from both the dynamic and static load cases are quite
low, even
when taking into account possibly higher stresses due to insufficient mesh
density at
stress concentrations.
[00145] All three dynamic loading cases resulted in separation of the
baseplate away
from the ground during the periodically upward portion of the load. The amount
of
deflection is small, but may increase for cases when the ground resonant
frequency is
excited (30 Hz). Further analysis with an excitation frequency of 30 Hz should
be
done to verify this. In no, loading case did the outer perimeter of the
baseplate leave
the ground, being held in place by the airbag load. This, too, should be
verified for
the 30 Hz case.
[00146] The separation from ground may be reduced or even eliminated by
judicious
design changes to stiffen the baseplate without penalizing weight, such as the
ribbed
configuration shown in Figure 14J.
EM Vibrator Cooling and Finite-Element Thermal Analysis
32
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
[00147] Three EMV components, which are field coils, armature lift coils, and
compensating plates, require active cooling to keep them from overheating
under
steady-state operating conditions. The heat losses generated in these three
components
under one embodiment of an EMV design are listed as follows:
= Field coil loss (12 pancake coils, 12 turns in each pancake coil)
- DC loss (calculated at 93 C) = 259 kW
= Armature lift coil loss
- DC loss in 8 lift-circuit conductors = 16 kW
- AC loss in other 56 lift-circuit conductors = 211 kW
- total loss = 16 kW + 211 kW = 227 kW
= Compensating plate loss
- AC loss (for 2 compensating plates) = 95 kW
= Total heat loss = 259 kW + 227 kW + 95 kW = 581 kW (779 hp)
[00148] The field coil and armature lift coil insulations are class-H
insulation material,
which can be continuously operated at 180 C. However, the potting resin used
to hold
the coil conductors and insulations together has a lower thermal rating. The
vibrator
cooling design need to assure that the maximum insulation temperature under
steady
conductor coil heat loads does not exceed the thermal degradation temperature
of the
potting resin. The coil insulation and potting resin details are described as
follows:
= Field coil and armature lift coil insulations
- conductor wrap: Austral VonRoll Isola 150 FCR Kapton' (0.0045 in.)
- armor coil wrap: medium weave glass tape (0.010 in.)
- class-H insulation (180 C)
= Potting resin
- PDG Y881K resin
- heat deflection temperature =120 C (yield strength)
- glass transition temperature = 131 C (failure strength)
[00149] As constrained by the heat deflection temperature of the potting
resin, the
cooling design goal of this machine is to keep the maximum copper coil
temperature
under 120 C under steady operations.
[00150] Since the ambient temperature varies from -35 C to 60 C, ethylene
glycol
water (40% ethylene glycol and 60% water by weight) has been chosen as the
coolant
to lower the freezing temperature and increase the boiling temperature of the
coolant.
33
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
Figure 15A shows the boiling and freezing temperatures of the ethylene glycol
aqueous solutions. The coolant inlet temperature to the machine is assumed to
be
65 C. Other cooling schemes and heat loads are contemplated within the scope
of
embodiments of the invention
[00151] The field coils are designed to be cooled by the coolant circulated
within both
copper conductor internal cooling passages and cooling passages attached to
the ends
of external cooling fins. Each field coil conductor is made of 0.25 in. square
copper
tubing with a 0.125-in hole in the middle to carry coolant. The thickness of
the copper
cooling fin is 0.1 in. The cooling tubes attached to the ends of each fin have
the same
size as that of the field coil conductors. A notional picture showing the
field coil
cooling design is shown in Figure 15B.
[00152] The armature force coils are cooled by the coolant circulated within
the
internal cooling passages in the copper conductors which are the same as those
of the
field coil copper tubing in terms of the dimensions. The compensating plate
heat loss
removal is accomplished by circulating the ethylene glycol coolant in four
circumferential cooling passages (0.5 in. diameter each) which are located in
the steel
surrounding the compensating plates. The armature lift coil and compensating
plate
cooling design is shown in Figure 15C.
[00153] To perform cooling analysis for predicting required coolant flow rate
and
coolant pumping power and to predict cooling boundary conditions required for
the
finite-element thermal modeling of the field coils, lift coils, and
compensating plates,
the coolant flow rates and cooling passage lengths need to be accurately
calculated for
all three cooling components. The input parameters used for the cooling
analysis are
listed in the following and some of the calculated cooling parameters are
included in
Table 2.
= Coolant flow rates
- in each pancake field coil cooling passage = 1.35 gpm
- in each field coil cooling fin cooling passage = 0.7 gpm
in each armature lift coil cooling passage = 1.2 gpm
in each compensating plate cooling passage = 6 gpm
= Numbers of coils or cooling passages
- pancake field coils= 12
- field coil cooling fin cooling passages = 10
34
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
- armature lift coils = 16
- compensating plate cooling passages = 4
= Component coolant flow rates
- field coil internal cooling passages = 16.2 gpm
- field coil fin cooling passages = 7 gpm
- armature lift coil cooling passages = 19.2 gpm
- compensating plate cooling passages = 24 gpm
= Total coolant flow rate
= 16.2 gpm + 7 gpm + 19.2 gpm + 24 gpm = 66.4 gpm
= Total cooling passage length
- each field coil internal cooling passage = 37.008 in
- each field coil cooling fin cooling passage = 3.025 in
each armature lift coil cooling passage = 12.649 in
each compensating plate cooling passage = 3.025 in
= Surface roughness = 0.00006 in. (copper tubing), 0.0018 in. (steel cooling
passage)
= Coolant pump efficiency = 80%
Table 2. Calculated cooling parameters
Field Coil Field Coil Armature Field Coil Field Coil Armature
Internal Fin Lift Coil Compensating Internal Fin Lift Coil Compensating
Component Passage Passage Passage Plate Passage Passage Passage Passage Plate
Passage
Section Average Average Average Average Exit Exit Exit Exit
Coolant Temp
( C) 89.71 84.16 89.37 73.28 114.42 103.32 113.73 81.56
Coolant
Velocity 10.95 5.65 9.73 3.00 11.17 5.74 9.92 3.02
(m/s)
Coolant
Convection
Coeff 38026 21568 34517 9160 45464 24896 41192 9731
(W/m2/K)
Coolant
Passage 37.01 3.03 12.65 3.03 37.01 3.03 12.65 3.03
Length
(m)
Coolant
Pressure 2348.44 59.22 647.83 4.93 2239.29 56.70 618.44 4.91
Drop
(psi)
Coolant
Pumping
Power 27.75 0.30 9.07 0.09 26.46 0.29 8.66 0.09
Required
(hp)
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
[00154] In Table 2, the energy balance calculations for field coil heat losses
removed
by coolant in internal cooling passages and coolant in external fin cooling
passages
were performed by assuming a 75%-25% heat removal split (75% heat loss removed
by internal cooling and 25% heat loss removed by external cooling). This 75%-
25%
split was obtained from thermal-circuit analysis. Material thermal properties
used in
the finite-element thermal analysis are included in Table 3.
Table 3. Material thermal properties
Material Density Specific Heat Thermal Conductivity
(kg/m3) (J/kg/K) (W/m/K)
Copper 8954 383 386
Insulation (after VPI) 1895 1129 0.19
Potting Resin 1300 1883 0.2
AISI 1005 Steel 7872 481 54
[00155] Finite-element meshes for two-dimensional axis-symmetric thermal
models of
field coils, lift coils, and compensating plates are shown in Figures 15D and
15E. Two
thermal analyses have been performed for each cooling component; one analysis
used
the cooling boundary conditions calculated at the average coolant temperature
and the
other analysis used the cooling boundary conditions calculated at the exit
coolant
temperature, which is the worst case. These cooling boundary conditions,
coolant
temperatures and coolant convection coefficients, are shown in Table 2. The
predicted
steady-state temperature distributions in the field coils, lift coils, and
compensating
plates are shown in figures 15F through 15K.
Testing
[00156] The test plan for the vibrator system can begin with component
development
tests early in the fabrication process. Eight force coils can be bonded to a
stainless
steel tube with the resin system proposed for the prototype unit. The mock-up
can be
faithful to the final design of the prototype force coils. After the mock-up
is cured it
will be instrumented in a press and the bond strength to the stainless tube
evaluated. A
pass criterion will be a bond strength at least a factor of two over that
required in
36
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
operation. The next component tests can be a steady state thermal test of a
water
cooled, insulated force coil and a water cooled, insulated field pancake coil.
The
appropriate electrical load that represents full force operation of the shaker
can be
applied to the coils. The coils can be heated until the temperature rise is
less than 2C
per hour. This will be considered steady state and the operating temperature
recorded.
The pass criterion for this test is that the steady state temperature does not
exceed the
heat deflection temperature of the resin system.
[00157] After the reaction mass is complete the cooling coils can be plumbed
to a hot
water source (149F) and a lab DC power supply can be used to pass operating
current
to the field coil. A Hall probe can be used in the air gap to the measure the
field
strength at several angular positions each time mapping the field as a
function of axial
position. An example of the data that can be collected is shown in Figure 16A.
A
passing inspection can be if the data is within 5% of the values predicted by
the FEA
magnetic analysis. During these tests the field coil temperature can be
monitored to
determine the steady state operating temperature of the as-built field coil.
The steady
state temperature should remain below the heat deflection temperature of the
resin
system.
[00158] When the base plate returns from Vacuum Pressure Impregnation (VPI) of
the
field coil it can be tested for correct frequency response. The baseplate can
be hung
from a cable and an instrumented with an accelerometer. An instrumented hammer
and data acquisition system can be used to find the eigenmodes of the
baseplate.
These can be compared to FEA structural models of the baseplate and should
match to
within 10%.
[00159] After rap testing the baseplate can be assembled with the reaction
mass and
the impedance of the force coils can be measured at discrete frequencies as a
function
of axial position in the reaction mass. An example of the data that can be
collected is
shown in Table 4. This data can be compared to the values predicted by the FEA
magnetics code and should match to within 10%.
Table 4. Force coil impedance as a function of frequency and position
x Test Test Test
Freq Freq Freq
20 Hz 50 Hz 120 Hz
-0.75 178 163 141
-0.25 73.5 73.7 67
37
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
0 54 53 53.1
0.25 55.8 55.1 53.3
0.5 71.2 69.2 67.2
[00160] With the vibrator assembled a series of DC tests can be performed. The
water
cooling and a DC power supply can be connected to the field coil. Additionally
a
second laboratory DC power supply can be connected to the water cooled force
coils.
The correct current can be established in the field coil and then the DC power
supply
connected to the force coil can be incremented until levitation is realized.
The test
levels can be compared to those predicted by the simulation. After the
levitation test
the baseplate can be locked to the reaction mass with a load cell. The force
coil can be
excited by itself and the reluctance force measured. This level can be
compared the
FEA magnetics analysis and recorded. The field coil can then be excited and
current
applied to the force coil in several positions established by shimming the
load cell.
The generated force can be recorded as a function of position.
[00161] At this point in the schedule the Pulsed Width Modulated (PWM)
Converter
can be delivered and checkout tests into a dummy load can be conducted.
Examples
of PWM test waveforms are shown in Figures 16B and 16C.
[00162] Once the PWM is deemed operational it can be connected to the shaker
and
low level vibration tests conducted. This allows debugging of the
instrumentation and
the data acquisition system. In addition the control gains may be adjusted for
optimal
operation. Once the vibrator has been demonstrated at low force integration
with the
truck can commence.
Power Mitigation Technologies
[00163] As it has been seen in previous sections if the vibrator was required
to produce
full force at 225Hz in steady state operation it would require 770 Hp of prime
mover
and the equivalent capability in heat rejection. In the prototype unit we
intend to
modify the Mertz engine to produce 600 Hp. The prototype can taper the force
at high
frequency to observe the 600 Hp rating of the diesel engine. For a production
unit, it
is desired to bring the required power down into the 600 Hp range while
maintaining
full force capability at 225 Hz.
[00164] There are several technologies that can reduce the power requirements
for the
Electromagnetic Vibrator (EMV) such as, for example, the use of permanent
magnets
in an embodiment of the EMV.
[00165] Another technology that can be used is spring assist. Springs can be
positioned
38
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
between the baseplate and the reaction mass to aid the Lorentz force (Figure
17A).
The first advantage seen is that springs can provide levitation force. Further
advantage
of springs can be seen by examining the shaker at its extremes of motion. In
Figure
17B it is seen that the springs are fully compressed in the full down
position. This is
when the vibrator needs the most help with up force. Gravity, the reluctance
force and
weakening of the Lorentz force all happen at this extreme of the displacement
range.
The springs are returning energy to the system right when it needs it the
most. The
other extreme in the range of motion is when the vibrator is in the full up
position
(Figure 17C). At this point of the stroke the spring has to be stretched and
is
receiving energy. It is at this point of the cycle that the baseplate current
and the field
current are in a direction to aid one another and establish more magnetic
field strength
in the air gap. The Lorentz force is at a maximum right when the spring is
stretched.
These observations point to a conclusion that the insertion of springs may
allow
reduction of both the field and armature currents to a level that would
provide
substantial reductions in power.
[00166] Another area for power savings involves conductor design. The eddy
current
and proximity losses may be lowered by reducing the exposed area in the
direction
that the time varying magnetic field intercepts the conductor. Figure 17D
shows a
design where the conductor size has been reduced from 0.25 in. to 0.125 in. In
addition the conductor has been transposed using a Roebel TM configuration to
reduce circulating currents (Figure 19-5). It can be seen that this conductor
change has
dropped the force coil losses from 306 kW to 249 kW.
[00167] While the methods and systems have been described in connection with
preferred embodiments and specific examples, it is not intended that the scope
be
limited to the particular embodiments set forth, as the embodiments herein are
intended in all respects to be illustrative rather than restrictive.
[00168] Unless otherwise expressly stated, it is in no way intended that any
method set
= forth herein be construed as requiring that its steps be performed in a
specific order.
Accordingly, where a method claim does not actually recite an order to be
followed
by its steps or it is not otherwise specifically stated in the claims or
descriptions that
the steps are to be limited to a specific order, it is no way intended that an
order be
inferred, in any respect. This holds for any possible non-express basis for
interpretation, including: matters of logic with respect to arrangement of
steps or
39
CA 02755664 2011-09-15
WO 2010/107759 PCT/US2010/027432
operational flow; plain meaning derived from grammatical organization or
punctuation; the number or type of embodiments described in the specification.
[00169] Throughout this application, various publications are referenced. The
disclosures of these publications in their entireties are hereby incorporated
by
reference into this application in order to more fully describe the state of
the art to
which the methods and systems pertain.
[00170] It will be apparent to those skilled in the art that various
modifications and
variations can be made without departing from the scope or spirit. Other
embodiments will be apparent to those skilled in the art from consideration of
the
specification and practice disclosed herein. It is intended that the
specification and
examples be considered as exemplary only, with a true scope and spirit being
indicated by the following claims or inventive concepts.
